First of all, the basic data covers a wide range, covering all visible and invisible information in the enterprise. Basic material information, product structure data, accounting subjects, supplier and customer information, department, factory, warehouse, workshop information and so on are all visible information, which will also be used in manual operation. Invisible information, such as document type, warehouse nature, planning parameters, etc. , will not involve manual information management, but will affect the system calculation.
In addition, the workload of basic data preparation is large, and the number of records of the above information ranges from several to hundreds of thousands, and each record can contain hundreds of fields. The product of the two is simply astronomical, and 90% of the reasons for the delay of the project usually come from the collation of basic data.
The correctness of the data is the most important. Basic data is the basis for many programs to run correctly. For example, the material plan and production plan are calculated according to the lead time, inventory and BOM structure set in the material file. Any data does not match the reality, and the planning result has no guiding significance.
It is precisely because of these characteristics of basic data that the collection and preparation work is heavy and the organization is difficult, which generally requires the coordination of multiple departments, and the input of manpower and time is relatively large, and the validity period is long, so the resistance is also great.
In order to help enterprises implement ERP more effectively, let's talk about how to prepare basic data quickly, at low cost and with low error rate.
Step 1: Determine the scope of work
Firstly, determine what data needs to be prepared according to the scope of ERP project, then determine the participating departments and staffing, and then determine the work plan. Remember not to hand over all the work to only one department or even one person, and be sure to fully realize the difficulty of this work. Regular meetings should also be arranged in the work plan to promote communication among staff.
Step 2: Establish the necessary coding principles.
The management of data by ERP software is realized by coding, which can uniquely identify data and run through future queries and applications. The establishment of coding principle is to make the follow-up work have a principle to follow, and also to determine the unique identification method that the database can identify for complex data. As the saying goes, sharpening a knife is not a mistake for a woodcutter, so we must not rush for success and ignore these important tasks.
In addition, the formulation of coding principles belongs to the formulation of enterprise-level standards and should be formulated and managed in accordance with ISO9000 standards, especially for a large number of basic data (such as the coding of material master documents), the scheme must be determined by multiple departments.
Step 3: Establish public information.
The established public information includes basic information such as company, subsidiary, factory, warehouse, department, employee information and currency code. These data will be referenced in other basic data, and the amount of data is not large, which can be completed with less time and manpower. If the lack of public information is found when sorting out other data, the overall efficiency and progress will be greatly reduced.
Step 4: Determine the BOM structure (optional according to the enterprise situation)
If an enterprise applies a production system, plan or product development module, BOM is the necessary basic data. First of all, we should make clear the hierarchical relationship between raw materials and semi-finished products, semi-finished products and products. The difficulty of this step is the setting of semi-finished products. If there are few or no levels of semi-finished products, the statistical analysis in the future cannot be refined; If there are many semi-finished products, the amount of data will increase greatly. In case of the following situations, the semi-finished products should be coded and managed: the semi-finished products inventory account should be established, or safety inventory management should be adopted, and the semi-finished products should be sold externally or used for after-sales service. In addition, semi-finished products do not need to be coded as far as possible and do not need to be entered into the software system. Every time a layer of BOM is added, the data volume of material information will increase accordingly. My personal opinion is that we can use as few BOM steps as possible to control the work.
Step 5: Collect first-hand information and original discrete data of different departments.
In these discrete data, there is only one basic material information, and the fields include production, purchase, sales, inventory and financial information. In this step, tables with uniform format should be used to cross and circulate among departments, so that each department can fill in data related to itself in the table, pass it to the next department after completion, and so on until this step is completed. Pay attention to the arrangement of transmission and determine the transmission path in advance. In order to ensure the construction period, different departments can start construction at the same time, and then cross traffic, or one department can complete a small part and then pass it on to the next department. At the same time, each distributed form is numbered uniformly (uniquely) and recorded during the handover between departments, which can not only control the progress, but also avoid data loss.
Step 6: Check the data
(1) Integrity check: Integrity refers to whether the number of records is complete. You can ask experienced people in the enterprise to review or calculate the total and compare it with historical data. At the same time, we should also check the integrity of the field. All ERP software has fields that must be entered. If these fields are missing, the system will be unstable, such as the lead time and default warehouse of the material. In addition, there are some fields that are not needed by the software, and fields that are useful for the future business and statistical analysis of the enterprise should also be included in the scope of investigation, such as customer classification and region.
(2) Correctness check: the range of correctness is very wide, and I don't need to explain them here. Enterprises can formulate inspection principles according to their own needs. Some mistakes, such as the account belongs to the asset category, but it is manually entered into the liability category, and some materials are purchased, but the self-made parts must be found and corrected before the system goes online.
(3) Uniqueness check: check the uniqueness of data from two angles. There are many common errors encoded in the same code. If it is entered into the system later, the mature ERP software will prompt that the code already exists and refuse to accept it. At the same time, the phenomenon that one physical object corresponds to multiple codes must also be eliminated. This kind of error can't be found by ERP software, so it must be found manually, otherwise there will be many accounting errors after going online.
At the same time of error correction, it is necessary to control the data version, which is particularly important in the data sorting work involving many departments. There was once a customer, and many departments were revising the same document at the same time. Only part of the data in the modified file is correct, and the data of other departments are still wrong. When inputting data, no matter whose data it is, it is unacceptable. You must replace the wrong data with the correct data and merge them. It can be seen that the responsible department and person in charge should be set for each type of data, and the person in charge will update the version of the document after each modification, and the old version of the data should also be saved. Doing so can minimize the loss of misoperation.
Step 7: Input the data into the software system.
Before input, the original files of basic data should be filed. For the data stored in electronic documents, backup should be made, and the sorting personnel, completion time and final version should be indicated. If it is printed paper media, it should be kept in a special filing cabinet and managed as an important document.
The next entry work is hard and boring repetitive work. Commonly used input methods are manual input and tool import. These methods have their own advantages and disadvantages.
(1) All softwares provide input interfaces, which can mobilize personnel to input data item by item, or use the copy function of the software to copy similar data, and then modify key fields. This method is relatively safe, at the same time, the perfect ERP software can automatically check the errors, and most of the errors not found in the early stage can be prompted by the system and prohibited from entering, thus ensuring the accuracy of the data. The disadvantage of this method is that the workload is large, and the completion time is multiple to the number of records. If the amount of data is not large, it is recommended to use this method, which is safe and reliable, and the workload of later inspection is also small.
(2) If there are database experts in the project team, and manual entry cannot be realized in terms of manpower and time, the method of directly importing the database can be adopted. The advantage of this method is high efficiency, and the completion time is multiple of the number of data tables. The disadvantage is that there are many hidden mistakes and it is not easy to check. Once a customer found a problem a few months after he went online, and after repeated checking, he found that a field in the data was vacant, but the entered data could be filled in automatically by the system. If this method is adopted, strict inspection steps should be set up in the early inspection work, especially in the correctness inspection, important fields should be checked one by one, and this work can be carried out manually with the help of various software technologies.
(3) The other method combines the advantages of the above two methods. Once, a customer developed a program to simulate manual work. The basic principle is to read the first field of the record from the spreadsheet (imitating the function of human eyes), then input data in the corresponding field in the ERP input interface (imitating the operation of human hands), read the next field again, and enter the program again. After this record is completed, go to the next record. This method uses the input interface of the program for input, and inherits the error function of interface input. At the same time, compared with data import, it can work 24 hours without manual intervention, and the speed is high, but it needs high-level programmers. So far, only one customer has successfully adopted this method.
Step 8: System Check
After completing the entry work, I still can't relax completely, so I must check it again. At this time, the best way is to test the data with software programs, such as backing up the database to a new database, doing the common processes in the new database again, and verifying the correctness of the basic data by checking the correctness of the results.
In the two steps of collection and collation, enterprises usually use some tools and software to assist data collection and collation according to the skills and specialties of employees, such as MS Excel, Access, FoxPro and SQL server. Using tool software can not only greatly improve the efficiency of collection and collation, but also has the greater advantage of being able to check errors. Most errors can be found by some simple instructions. If we rely on manual comparison and inspection, the time and workload required are unbearable for enterprises.
The basic data collation process of enterprise resource planning project was discussed above. Specific projects should be carried out under the guidance of the implementer's consultant, and the system guarantee should be kept up in time. Before data collation, some enterprises adopted the suggestions of implementers and established a reward and punishment system. Input data is calculated at 0.02 yuan, and 0.04 yuan will be deducted if the input data is found to be incorrect after inspection. A system not only improves the enthusiasm of the input personnel, but also urges the input personnel to conduct a serious inspection.
The focus of ERP basic data preparation is not the data itself. Strict planning and reasonable organization are the best means to complete this arduous work.